teach
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from multiple sources within the workspace and external resources, which creates a potential surface for indirect prompt injection.
- Ingestion points: Workspace files including
MISSION.md,RESOURCES.md, and./learning-records/*.md, as well as external content from suggested high-reputation communities and resources. - Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within these data sources.
- Capability inventory: The agent can write to various files and directories (
./lessons/,./assets/,NOTES.md, etc.) and is instructed to run CLI commands to open generated files. - Sanitization: There are no explicit instructions for the agent to sanitize or validate the external content it processes.
- [DYNAMIC_EXECUTION]: The skill framework requires the agent to generate interactive web components (such as JavaScript simulators and widgets) and store them in the
./assets/directory. These components are then executed in the user's browser, representing a form of dynamic code generation. - [COMMAND_EXECUTION]: The instructions direct the agent to run CLI commands to open the generated lesson files for the user's review.
Audit Metadata